AI Engineer
Sign up free to see how well your resume matches this role.
What you'll do
- Building investigation agent pipelines (planning, tool calling, synthesis).
- Designing and maintaining an evaluation framework to measure system performance and track metrics over time.
- Evaluating foundation models (open-source and proprietary) against our use cases and documenting trade-offs between cost, latency, and quality.
- Industrializing developments within a shared codebase through testing, packaging, documentation, and CI/CD best practices.
- Integrating solutions into the product, deploying them, and monitoring them in production.
What they're looking for
- You can read and understand Python code.
- You have a solid understanding of how LLMs work, including data gathering, prompting, fine-tuning, inference, attention mechanisms, and tokenization.
- You have hands-on experience working with LLMs through APIs or local deployments and have built applications such as RAG systems, tool-calling workflows, agents, or chatbots.
Nice to have
- Basic knowledge of cybersecurity.
- Experience with agent frameworks such as LangGraph, PydanticAI, Haystack AI, or similar.
- Knowledge of machine learning or deep learning frameworks such as Scikit-learn, PyTorch, or equivalent.
Summarised by NextRaise from the employer’s description, which follows in full below.
Full description from employer
Who we are?
HarfangLab is a French cybersecurity scale-up developing a unified platform powered by a single lightweight agent and AI built for cybersecurity. Continuously enhanced through R&D, our portfolio includes EDR (Endpoint Detection and Response), EPP (Endpoint Protection Platform), ASM (Attack Surface Management), ITDR (Identity Threat Detection and Response), and Email Security, with the latter two joining the platform in 2026. Together, these solutions detect and neutralize threats across endpoints, servers, identities, and email environments. Our engines, models, and algorithms identify abnormal behaviour, generate security alerts, and block malicious activity.
Founded in 2018, HarfangLab has earned strong industry recognition: its EDR was the first to be certified by ANSSI in 2020, the first to obtain ANSSI qualification in 2025, and the first EDR to receive BSI certification in Germany. We also joined the French Tech 2030 programme, alongside 80 innovative French companies, as cyber resilience becomes a strategic priority for Europe.
Backed by €30 million in funding, we have grown from 40 employees in 2023 to 140 today and expanded internationally across Europe.
HarfangLab now protects the endpoints of more than 1,000 clients across all sectors, including CAC 40 industrial companies, government entities, SMEs, hospitals, and local authorities. Driven by a strong collective spirit and international ambitions, we remain committed to our original mission: protecting organizations against increasingly complex cyber threats while enabling them to retain full control over their data.
What you will do with us?
You will join the Artificial Intelligence team, a group of 7 people. This team has developed an initial system leveraging a Large Language Model to help cybersecurity analysts identify and respond to threats. The team is now working on enhancing this system with new capabilities, enabling it to perform advanced attack investigations autonomously.
You will contribute to the design, development, and deployment of these new features, from prototype to production. Your responsibilities will include:
Building investigation agent pipelines (planning, tool calling, synthesis).
Designing and maintaining an evaluation framework to measure system performance and track metrics over time.
Evaluating foundation models (open-source and proprietary) against our use cases and documenting trade-offs between cost, latency, and quality.
Industrializing developments within a shared codebase through testing, packaging, documentation, and CI/CD best practices.
Integrating solutions into the product, deploying them, and monitoring them in production.
Depending on your interests and background, you may also have the opportunity to work on machine learning, deep learning, or data engineering projects.
About You
Hard Skills:
You can read and understand Python code.
You have a solid understanding of how LLMs work, including data gathering, prompting, fine-tuning, inference, attention mechanisms, and tokenization.
You have hands-on experience working with LLMs through APIs or local deployments and have built applications such as RAG systems, tool-calling workflows, agents, or chatbots.
Nice to have:
Basic knowledge of cybersecurity.
Experience with agent frameworks such as LangGraph, PydanticAI, Haystack AI, or similar.
Knowledge of machine learning or deep learning frameworks such as Scikit-learn, PyTorch, or equivalent.
Soft Skills:
You are curious and autonomous, and you are comfortable taking ownership of topics that are not yet fully defined.
You have strong analytical skills and can help structure complex problems so that other team members can contribute effectively.
You enjoy collaborative work, communicate clearly, and are comfortable asking for or offering help when needed.
You are motivated to actively contribute to integrating your work into the HarfangLab product.
What We Offer
Our Office and Team Culture
Offices located in the heart of Paris, near Bourse (75002),
High-performance equipment tailored to your needs and preferences (Mac/PC, additional screens),
Thanks to our Office Manager, events regularly take place such as seminars, themed evenings, and more,
An onboarding process to welcome each new colleague with an explanation of the roles within the different teams of the company and a mentor to guide you during your early days.
Compensation & Benefits
Total compensation: €44-54K + profit sharing
Health, well-being & sport: Alan insurance + Alan Mind mental wellness app + Egym subscription, covered at 80% by HarfangLab,
Meal vouchers: Swile card + access to an employee benefits platform.
Paid time off: 25 days annual leave + 8-11 RTT days.
Professional growth: Access to training programs and industry events tailored to your needs.
Company
Company facts come from this company's own listings. We only show what the postings themselves carry.